The Biosynthesis of the Thiazole Phosphate Moiety of Thiamin: the Sulfur Transfer Mediated by the Sulfur Carrier Protein This
Total Page:16
File Type:pdf, Size:1020Kb
Chemistry & Biology, Vol. 11, 1373–1381, October, 2004, 2004 Elsevier Ltd. All rights reserved. DOI 10.1016/j.chembiol.2004.08.009 The Biosynthesis of the Thiazole Phosphate Moiety of Thiamin: The Sulfur Transfer Mediated by the Sulfur Carrier Protein ThiS Pieter C. Dorrestein, Huili Zhai, ThiS-carboxylate. This was unambiguously shown using Fred W. McLafferty, and Tadhg P. Begley* high resolution electrospray ionization Fourier mass Department of Chemistry and Chemical Biology spectrometry (ESI-FTMS) of the undigested 13C- and Cornell University 15N-depleted protein formed from 18O-labeled DXP. In Ithaca, New York 14853 addition, we have trapped and identified a new DXP- derived thioenolate intermediate covalently linked to ThiG. We propose a mechanism for the complex biosyn- Summary thesis of the thiazole-phosphate moiety based on these findings. Thiamin-pyrophosphate is an essential cofactor in all living systems. The biosynthesis of both the thiazole Results and Discussion and the pyrimidine moieties of this cofactor involves new biosynthetic chemistry. Thiazole-phosphate syn- The thiazole moiety 8 (Figure 1) is biosynthesized in thase (ThiG) catalyses the formation of the thiazole Bacillus subtilis and most other bacteria from 1-deoxy- moiety of thiamin-pyrophosphate from 1-deoxy-D- D-xylulose-5-phosphate (1, DXP), glycine, and cysteine xylulose-5-phosphate (DXP), dehydroglycine and the in a complex oxidative condensation reaction [5]. This sulfur carrier protein (ThiS), modified on its carboxy reaction requires five different proteins, ThiO, ThiG, terminus as a thiocarboxylate (ThiS-thiocarboxylate). ThiS, ThiF, and a cysteine desulfurase. Glycine oxidase Thiazole biosynthesis is initiated by the formation of a (ThiO) catalyzes the oxidation of glycine to the corre- ThiG/DXP imine, which then tautomerizes to an amino- sponding imine 7, sulfur carrier protein adenylyl trans- ketone. In this paper we study the sulfur transfer from ferase (ThiF) catalyzes the adenylation of the carboxy ThiS-thiocarboxylate to this amino-ketone and trap terminus of the sulfur carrier protein (ThiS-carboxylate), a new thioenolate intermediate. Surprisingly, thiazole and cysteine desulfurase catalyzes the transfer of sulfur formation results in the replacement of the ThiS-thio- from cysteine to the ThiS-acyl adenylate to give ThiS- carboxylate sulfur with an oxygen from DXP and not thiocarboxylate (6) [5, 8, 9, 19]. ThiG is the thiazole syn- from the buffer, as shown by electrospray ionization thase and catalyzes the formation of the thiazole from Fourier transform mass spectrometry (ESI-FTMS) us- dehydroglycine 7, DXP 1, and ThiS-thiocarboxylate 6. ing 18O labeling of the 13C-, 15N-depleted protein. These The early steps in thiazole formation have been eluci- observations further clarify the mechanism of the dated [10]: Imine formation between lysine 96 on ThiG complex thiazole biosynthesis in bacteria. and DXP followed by tautomerization gives amino- ketone 5, which is then proposed to react with ThiS- Introduction thiocarboxylate 6 and dehydroglycine 7 to give thiazole phosphate 8. Thiamin pyrophosphate (9, Figure 1) is an essential co- During the formation of thiazole-phosphate (8), inter- factor in all living systems and consists of a pyrimidine mediate 10 formed by the addition of 6 to the C3 of 5, covalently linked to a thiazole. This cofactor is utilized could undergo hydrolysis, releasing ThiS-carboxylate, for reactions catalyzed in amino acid metabolism, the followed by loss of water to give thioketone 12 (mecha- pentose phosphate pathway, and the citric acid cycle. nism A, Figure 2). Enolization of 12 would give 13, which [1, 2] Furthermore, thiamin-triphosphate has been impli- could react with the dehydroglycine 7 to give the thiazole cated in brain disorders [2]. In thiamin deficient humans, 8 or eliminate water to give 14, which could then react these processes do not function properly and result in with the dehydroglycine. Alternatively, sulfur to oxygen disease states known as Beri-Beri or Werninke-Korsa- acyl shift in 10 would give 15, loss of water would give koff syndrome, both of which can be fatal [2]. Recently thioketone 16, which could then enolize to give 17 (2003), fatal Beri-Beri was diagnosed in infants from (mechanism B, Figure 2). This could react with the dehy- Jewish communities in Israel as a result of the consump- droglycine 7 to give the thiazole 8 or eliminate ThiS- tion of thiamin-deficient baby-food products [4]. Consid- carboxylate to give 14, which could in turn react with the ering the important cellular roles thiamin plays, it is sur- dehydroglycine. The experiments reported here allow us prising that we are now just beginning to understand to differentiate between mechanisms A and B. its biosynthesis in bacteria, while its biosynthesis in eu- karyotes is still at an early stage [3, 5–7]. In this paper, we Determination of the Source of the Carboxy- characterize the mechanism of formation of the thiazole- Terminal Oxygen of ThiS-Carboxylate phosphate moiety (8) of thiamin in vitro starting with the Derived from 6 by ESI-FTMS sulfur transfer reaction from ThiS-thiocarboxylate to the One can differentiate between mechanisms A and B in amino ketone 5, during which a hydroxyl group from Figure 2 by determining the origin of the oxygen on ThiS- 1-deoxy-D-xylulose-5-phosphate (1) (DXP) is trans- carboxylate that replaced the sulfur of ThiS- thiocarbox- ferred to the C-terminal end of the sulfur carrier protein, ylate (6). For mechanism A, this oxygen will be derived from the buffer, whereas for mechanism B, it will be *Correspondence: [email protected] derived from DXP. Chemistry & Biology 1374 Figure 1. The Biosynthesis of the Thiazole- Phosphate Moiety (8) of Thiamin Pyrophos- phate (9)inB. subtilis To determine the origin of this oxygen, ThiS-thiocar- species. 13C-, 15N-depleted ThiS-carboxylate was enzy- boxylate (6) and DXP (1) were incubated with ThiG in matically converted to 13C-, 15N-depleted ThiS-thiocar- 18O buffer and the resulting ThiS-carboxylate was ana- boxylate. The mass spectrum of this protein is shown lyzed by ESI-FTMS to give the spectrum shown in Figure in Figure 4C and shows the expected ϩ16 Da mass 3B. The observed molecular ion was 10146.4 Da, identi- increase. When 13C-, 15N-depleted ThiS-thiocarboxylate cal to the calculated mass of ThiS-carboxylate (Figure and [3,4-18O]-DXP (prepared using dihydroxyacetone- 3C), suggesting that 18O from buffer was not incorpo- phosphate, pyruvate, triose-phosphate isomerase, and rated into ThiS-carboxylate. When a similar reaction was DXP-synthase in 80% 18O-buffer) were incubated with run in 16O buffer using partially labeled [2,3,4-18O]-DXP ThiG in 16O-buffer, the resulting 13C-, 15N-depleted ThiS- (prepared using dihydroxyacetone-phosphate, pyruvate, carboxylate had a measured mass of 10142.3 Da, 2 Da 13 15 triose-phosphate isomerase, and DXP-synthase in 70% larger than C-, N-depleted ThiS-carboxylate (Figure 18O-buffer [10, 11]), the corresponding molecular ion had 4D), the calculated mass for which is 10142.1 Da. The 18 a mass between 10146.4 and 10148.4 Da (Figure 3A), amount of O-label incorporated into ThiS-carboxylate suggesting that the new carboxy-terminal oxygen of in this experiment was estimated to be 78% (data not ThiS-carboxylate was derived from DXP rather than from shown). In contrast, an identical reaction in which [3,4- 18O]-DXP was replaced with DXP, without 18O-labeled the buffer. We have obtained further support for this oxygens, gave 13C-, 15N-depleted ThiS-carboxylate with conclusion using 13C-, 15N-depleted ThiS-thiocarboxy- the mass spectrum shown in Figure 4E, similar to the late as described below. starting 13C-, 15N-depleted ThiS-carboxylate shown in Figure 4B. SWIFT isolation and IR-multiphoton dissocia- Confirmation of the Source of the Carboxy- tion of the reformed ThiS-carboxylate ion observed in Terminal Oxygen Derived from 6 by ESI-FTMS Figures 3A and 4D resulted in fragmentation at 22 differ- 13 15 Using C-, N-Depleted ThiS-Thiocarboxylate ent sites. None of the b fragments (amino terminal) con- To simplify the mass spectrum of ThiS-carboxylate taining residues 84, 85, and 86 of the 87 amino-acid- shown in Figure 3A, we were able to reduce the ThiS- reformed ThiS-carboxylate carried the 18O label (ϩ2Da carboxylate isotopic cluster to a single major species modification) while all the y fragments (carboxy terminal) using 13C-, 15N-depleted ThiS-carboxylate [12]. To ac- did (Figures 4F and 4G). While this experiment does not complish this, 13C-, 15N-depleted ThiS-carboxylate was exclude the possibility that the oxygen replacing the overexpressed and purified from minimal medium con- sulfur of ThiS-thiocarboxylate is derived from the C3 taining 13C-depleted glucose and 15N-depleted ammo- rather than the C4 oxygen of DXP, the experiment unam- nium sulfate. The mass spectrum of this protein is shown biguously demonstrates that the oxygen of the reformed in Figure 4B and is compared to the mass spectrum of ThiS-carboxylate is derived from DXP and not from the native ThiS-carboxylate in Figure 4A. The monoisotopic buffer. Therefore mechanism A in Figure 2 can be ex- ion, with an observed mass of 10140.3 Da, is the major cluded from further consideration. Figure 2. Mechanistic Analysis of the Middle Steps in the Formation of Thiazole Phosphate Thiamin Thiazole, Sulfur Transfer Mechanism 1375 NaBH4 and the resulting ThiG was analyzed by ESI- FTMS for covalent modification (Figure 5D). The mass of the resulting ThiG molecular ion (27004 Da) was 200 Da larger than that of native ThiG (26804 Da). This is not consistent with the trapping of 20 because the mass of reduced 20 is ThiG ϩ 184 Da. The observed mass in- crease of ϩ200 Da is consistent with reduced 5 (Figure 6A) and with reduced 14.